Discussion
Approximately one in three adolescents experiences elevated depressive symptoms (Shore et al.,
2018), and some of them also report short- or long-term negative consequences (Weavers et al.,
2021). Emotion regulation has been extensively studied as a proximal factor associated with depressive symptoms. However, most of these studies have focused on only a limited number of strategies. Moreover, the majority of research has been cross-sectional, especially in the context of examining the protective effects of adaptive strategies. Hence, this study aims to investigate how various emotion regulation strategies, including both adaptive and maladaptive ones, predict depressive symptoms over time. The longitudinal design of the study also enables to explore differences in emotion regulation strategies among subgroups of adolescents based on their patterns of depressive symptoms over time. Generally, this study found that both maladaptive and adaptive emotion regulation strategies predicted depressive symptoms cross-sectionally and prospectively in adolescents. While the concurrent association between maladaptive emotion regulation strategies and depressive symptoms was stronger compared to the concurrent association between adaptive emotion regulation strategies and depressive symptoms, the adaptive strategies had comparable effect on depressive symptoms prospectively. In addition, three subgroups were identified based on the intensity of depressive symptoms across waves which differed from each other on both the putatively adaptive and maladaptive strategies measured at Time 1.
Existing research, both cross-sectional (Schafer et al.,
2017) and prospective (Cavicchioli et al.,
2023), suggests that inter-individual differences in emotion regulation explain depressive symptoms during adolescence, and the current study supports this well-established association. As stressors and the context in which emotions are regulated can considerably vary, the efficacy and adaptiveness of certain strategies may also differ (Sheppes & Meiran,
2007). Therefore, it is not surprising that most strategies loaded positively on both factors in the analysis. However, the magnitude of factor loadings reflected the supposed nature of the strategies. For example, positive reappraisal had a considerably higher factor loading on the adaptive emotion regulation factor (0.75) than on the maladaptive emotion regulation factor (0.07). Conversely, rumination had a stronger loading on the maladaptive factor than on the adaptive factor (0.71 vs. 0.11). The stark contrast in factor loadings in these cases suggests that reinterpreting stressful events in a more positive light (as defined in the CERQ) is a more adaptive approach, while dwelling on stressful events through rumination is a less adaptive one. However, there are situations where reinterpreting an event or situation in a positive manner can hinder the attainment of long-term goals (Aldao et al.,
2015). For example, consistently perceiving additional work as a sign of trust could negatively impact one’s ability to spend quality time with loved ones and, in turn, have an adverse effect on mental health. Similarly, from some perspectives, ruminating over events that indicate a discrepancy between one’s desired and actual state is a normal process, provided it is time-limited (Koster et al.
2011). Interestingly, positive refocusing, which is akin to distraction, had a negative loading on the maladaptive factor and a positive loading on the adaptive factor, suggesting that individuals who engage in self-blame, rumination, or catastrophizing may be unable to divert their attention from negative stimuli, both internal and external. In many cases, diverting attention from negative events, such as a classmate being unkind, is beneficial because it prevents individuals from falling into a negative spiral. However, the low factor loadings of positive refocusing on the adaptive and maladaptive factors (+0.311 and −0.268, respectively) suggest that the adaptiveness of this strategy is highly context dependent. For example, if positive refocusing or distraction serves as a means of avoiding internal experiences, a phenomenon referred to as experiential avoidance (Hayes et al.,
1996), it is more likely to become maladaptive, as it could exacerbate negative emotions. On the other hand, catastrophizing is likely to be maladaptive regardless of the context, as it had a negative loading on adaptive strategies (−0.18) and a positive loading on maladaptive strategies (0.71).
Consistent with previous meta-analytic results in studies involving adults or adolescents (Aldao et al.,
2010; Schafer et al.,
2017), the current study also found that maladaptive emotion regulation strategies are strongly associated with depressive symptoms concurrently. However, the main finding is that the prospective association between adaptive emotion regulation strategies and depressive symptoms is comparable to that of maladaptive emotion regulation strategies and depressive symptoms. Therefore, not only the use of maladaptive emotion regulation strategies but also the reduced use of adaptive emotion regulation strategies (reappraisal, putting into perspective, positive refocusing, and acceptance) during stress is an equally important prospective predictor of depressive symptoms. In other words, an increased utilization of these strategies is linked to a reduction in depressive symptoms among youth. The negative relationship between the use of adaptive emotion regulation strategies and depressive symptoms has predominantly been established through cross-sectional studies (Kraft et al.,
2023). However, there is a scarcity of longitudinal studies that explore the connection between depressive symptoms and the employment of adaptive strategies (Cavicchioli et al.,
2023). Regarding adaptive emotion regulation strategies, it is noteworthy that studies typically measure acceptance, reappraisal, or problem-solving. However, in the present study, rather than measuring problem-solving, we assessed planning, defined as a series of actions one can take to either change or cope with a situation (Garnefski et al.,
2001). However, as the definition implies, planning alone does not guarantee problem resolution. The planning process may not necessarily result in a final decision on the course of action or even the implementation of a potential solution. This might explain why planning did not load onto either the adaptive or maladaptive factor.
Repairing mood with different strategies, such as reinterpreting the events or refocusing attention on positive things, could be protective factors. For instance, a longitudinal study among college students found a prospective link between cognitive reappraisal, close peer relationship, and better sociometric position (English et al.,
2012). Thus, it is possible that reappraisal protects against mental health problems through improved social functioning. However, it is also possible that a higher level of depressive symptoms may result in less successful reappraisal (Bettis et al.,
2019). This notion is supported by a meta-analysis of 13 fMRI studies (mainly with adult participants) that found that depressed patients exhibit reduced activation in the cognitive control network during cognitive reappraisal tasks compared to healthy controls, which suggests that reappraisal may not be effective in depression (Picó-Pérez et al.,
2017). Similarly, a study found no relationship between reappraisal success and activity in the dorsolateral and dorsomedial prefrontal cortex in depressed adolescents, which further supports the idea that reappraisal may not be effective in this population (LeWinn et al.,
2018).
Furthermore, the choice of a particular strategy is dependent on the situation. For instance, when compared to distraction, cognitive reappraisal is preferred in low-intensity situations according to laboratory studies using strategy choice paradigm (Van Bockstaele et al.,
2020). This may be due to it being easier to reinterpret mild negative situations compared to events that trigger strong negative emotions. However, there was no relationship between reappraisal and emotional intensity in an ESM study investigating emotion regulation strategy use in naturalistic settings among adolescents. Instead, that study found that adolescents chose rumination, relaxation, or emotional expression when they experienced intense negative emotions (De France & Hollenstein,
2022). In the context of depression, although depressed people are able to use reappraisal when instructed to do so (e.g., following sadness induction in the lab), they do not appear to choose this strategy spontaneously for reducing their sadness (Ehring et al.,
2010). In fact, they usually choose strategies that maintain or even increase their negative mood (Quigley & Dobson,
2014). As reappraisal is related to increased positive affect concurrently and prospectively (lagged-effect) in ESM and daily diary studies (Boemo et al.,
2022), a relative lack of reappraisal in everyday life may lead to decreased positive emotions, which in turn could contribute to depressive symptoms. Or alternatively, ineffective use of reappraisal in depression may prevent long-term benefits of this emotion regulation strategy.
Putting into perspective is also considered a kind of appraisal tactic, as it provides a new perspective for evaluating the event. Namely, this strategy refers to comparing the current situation to something worse (e.g., “I tell myself that there are worse things in life”) (Garnefski et al.,
2001), thus it is reminiscent of downward counterfactual thinking (e.g., much worse things could have happened) (Davey & McDonald,
2000). Putting into perspective does not require generating new meaning, therefore it may be easier to apply than positive reappraisal in adolescence. However, putting into perspective is used less often by depressed young people than their non-depressed peers in a stressful situation (van den Heuvel et al.,
2020).
Various studies have shown the beneficial effects of accepting emotions (Lennarz et al.,
2019) and the negative effects of non-acceptance of emotions (including experiential avoidance) (Lydon-Staley et al.,
2019; Schafer et al.,
2017; Szemenyei et al.,
2020). However, the CERQ does not measure this, but the degree of acceptance of the stressful situation itself which is sometimes labelled as resignation (Wolgast et al.,
2013). But even so, the CERQ acceptance subscale was more strongly loaded on the adaptive factor in the current study, suggesting that there could be situations in which passively accepting the situation could be an adequate response to stress in adolescents. One study (Lennarz et al.,
2019), for example, found that adolescents used acceptance most often in everyday life, and they also found situations with less intense emotions easier to accept. It is worth noting, however, that this research only measured emotion regulation on weekend days in a normative sample. In clinical samples, acceptance appears to be more maladaptive. For example, acceptance of stressful situations (as measured by the CERQ) was positively associated with symptoms among adolescents diagnosed with major depressive disorder (Ding et al.,
2021). Based on these results, acceptance of certain situations may be particularly beneficial for mental health if they are otherwise difficult to control and change, in which case acceptance may act as a form of secondary coping (Compas et al.,
2001). However, when used in controllable context as a form of resignation, it may work against problem solving.
Although two data points do not allow for trajectories, the latent class analysis method can show some patterns of symptom evolution even with two data points. The analysis indicated that the three groups that emerged differed in the degree of symptom prevalence (low to medium to high), and no groups with increasing or decreasing symptoms were found. In line with previous studies (Shore et al.,
2018), most of the students belonged to the stable low symptoms group (approximately 60% of the students), but a substantial proportion of youth (around 30%) reported somewhat elevated symptoms occurrence at both waves (stable moderate symptoms group). Importantly, around 30% of the students were part of the moderate symptoms group, with a mean score around the suggested cutoff (24, Rushton et al., 2022; Shore et al.,
2018) for the CES-D 20 item questionnaire. A small percentage (around 7%) of youth, however, were in the stable high symptoms group. These proportions are consistent with the results of a meta-analysis (Shore et al.,
2018) or some studies using the CES-D to measure depressive symptoms (Ferro et al.,
2015). Conversely, there are some studies (e.g., Rodriguez et al.,
2005; Yaroslavsky et al.,
2013) that identified a higher proportion of students with stable high symptoms using the CES-D. In line with previous studies (Shore et al.,
2018), the current study also found that in both the stable moderate and stable high symptoms groups there were more girls than in the stable low symptoms group.
High symptom groups were reliably replicated across studies (Shore et al.,
2018), suggesting that for some youth experiencing high depressive symptoms are relatively stable over time. In the current study, the use of maladaptive emotion regulation strategies (rumination, self-blame, and catastrophising) at Time 1 was more pronounced in this group compared to stable low depressive symptoms group. However, increased use of maladaptive emotion regulation strategies was also present when comparing the stable moderate symptoms group to the stable low symptoms group. The maladaptive strategies measured in this study share the common feature of focusing on the negative aspect of the self, and therefore, could easily co-occur. For instance, self-critical or self-blaming ruminative thoughts can easily lead someone to catastrophize; thus, eliciting them to perceive the situation as much worse than it really is. Biased attention towards, and the increased awareness of, negative external or internal stimuli are both risk and maintenance factors of depression (Disner et al.,
2011).
Maladaptive emotion regulation strategies may also contribute to the maintenance of depressive symptoms through other mechanisms. For instance, in a short prospective study excessive reassurance seeking as a maladaptive interpersonal behaviour mediated the association between emotion regulation difficulties and high depressive symptoms (Fearey et al.,
2021). Rumination can prolong the emotional or physiological concomitants of stressors even after the stressors have terminated (Radstaak et al.,
2011); thus, rumination may contribute to depression via transforming past stressors into present events, in turn reinforcing rumination (Hosseinichimeh et al.,
2018). Interestingly, this latter mechanism was found to be stronger in adolescent girls compared to boys (Hosseinichimeh et al.,
2018), suggesting that not only increased rumination among girls (Hankin,
2008) but also gender differences in rumination-related mechanisms might contribute to the well-known gender differences in depression.
Clinical Implications
Overall, based on the proportion of adolescents reporting moderate or high levels of symptoms, monitoring depressive symptoms during adolescence would be important. According to a relatively recent meta-analysis, more than a third (34%) of adolescents aged 10–19 years suffer from mild, albeit often subclinical/subthreshold level of depression (Shorey et al.,
2022). A proportion of the group with even moderate symptoms in the current study are likely to meet the criteria for subthreshold depression. Subthreshold depression is a clinically significant condition (Bertha and Balazs,
2013; Noyes et al,
2022), the onset and persistence of which are poorly understood in adolescents.
The findings also suggest that young individuals with moderate to high depressive symptoms may benefit more from intervention programs that focus on emotion regulation. For instance, in the current study students with moderate symptoms had over three times higher odds of engaging in maladaptive emotion regulation strategies compared to those in the low symptom group, whereas students in the high symptom group had more than eight times higher odds of using such strategies. Thus, interventions that help adolescents with moderate to high symptoms to recognize and reduce the use of maladaptive emotion regulation strategies and to adopt more adaptive methods to manage negative emotions are warranted in non-clinical settings as well. For example, cultivating decentering (a metacognitive capacity to observe inner experiences from a distanced perspective) at this stage of development, where there is heightened self-awareness and intense negative emotions, may be effective in ameliorating distress and reducing perseverative thoughts (such as rumination) (Bennett et al.,
2021).
Improvements in interventions aimed at regulating emotions within clinical populations are typically associated with decreased psychopathological symptoms (Moltrecht et al.,
2021). This suggests that positive changes are feasible in community samples with moderate to high symptoms, if baseline levels of various psychopathological symptoms are considered in these effectiveness trials. Furthermore, the limited impact of interventions designed to enhance emotion regulation in community samples (Eadeh et al.,
2021) may be partly attributed to a significant portion of young individuals who do not face substantial challenges in regulating their emotions. Thus, a floor effect may exist within this population (Eadeh et al.,
2021).
Limitations
There are some limitations to the current study. This study only examined urban adolescents in a narrow age range; therefore, generalizability to rural and minority adolescents is limited. As emotion regulation was only measured at the first time point, the scar effect hypothesis, i.e., how increased depressive symptoms affect the use of emotion regulation strategies within individuals across time, could not be tested. A meta-analysis on longitudinal studies in children and adolescents (Cavicchioli et al.,
2023) showed that after using the cross-lagged panel model, both adaptive and maladaptive emotion regulation strategies at Time 1 were significant predictors of psychopathological symptoms at Time 2. However, the relationship was mutual due to psychopathological symptoms at Time 1 also predicting adaptive and maladaptive emotion regulation at Time 2. As emotion regulation strategies were not measured at Time 2, mutual relationship could not be tested in the current study. Additionally, only six months passed between the two data collections; therefore, the long-term prediction of symptoms through emotion regulation remains uncertain.
The questionnaire (CERQ) used generally measures emotion regulation strategies in the presence of a stressor. Accordingly, the current study lacked the information on how different characteristics of a given stressful situation (or a life event), such as the intensity of emotions associated with it and its controllability, influence the use of emotion regulation strategies (De France & Hollenstein,
2022; Lennarz et al.,
2019). Additionally, it should be noted that the CERQ does not assess particular strategies, including expressive suppression, behavioural or experiential avoidance, which research has linked to psychopathological symptoms in adolescents (Kraft et al.,
2023). Regarding avoidance, it has been proposed that dissociation could potentially be an automatic mechanism that contributes to the avoidance of emotional experiences (Cavicchioli et al.,
2021).
There are also some findings suggesting that emotion regulation should be assessed in an emotion-specific manner (De France & Hollenstein,
2022). Furthermore, using a specific strategy does not guarantee the success of regulation. It can happen that even after using adaptive strategies, the emotion remains dysregulated (Jazaieri et al.,
2013), which cannot be measured by administering the CERQ.
Current study focused on stress-related emotion regulation, but a lack of ability to generate and maintain positive emotions (Nelis et al.,
2015) or dampening positive emotions has also been found to be related to depressive symptoms (Bean et al.,
2022; Burke et al.,
2018). Furthermore, taking into account that emotion regulation can take place not only intrapersonally, but also in social contexts, suggests that investigating co-regulation patterns can add to the understanding of mental health in adolescence. For instance, co-rumination (excessive talking and dwelling) on negative emotions (Rose,
2002) with friends/peers has been found to be associated prospectively with depressive symptoms (Bastin et al.,
2018) possibly via increased intrapersonal brooding rumination (Bastin et al.,
2021).
The study focused on stress-related emotion regulation, but perceived stress was not assessed. However, everyday academic stress and interpersonal conflicts with peers and family members are consistently associated with negative daily mood (Bai et al.,
2017). Difficulties in effectively altering negative mood or emotions triggered by these everyday events might contribute to depressive symptoms over time (Kovacs & Yaroslavsky,
2014). For instance, an ecological momentary assessment study conducted among adolescents (Rothenberg et al.,
2019) found that increased daily sadness is associated with increased daily levels of depressive symptoms when the adolescent has more difficulty regulating their emotions on that day. Not only perceived stress as a proximal factor but also negative adverse childhood experiences as distal factors can lead to the development of depressive symptoms and increased levels of rumination later in life (Mansueto et al.,
2021). However, in the present study, we did not assess distal factors, including adverse childhood experiences.
Conclusion
Emotion regulation strategies have been extensively studied in relation to psychopathology or symptoms. However, empirical studies have predominantly focused on maladaptive strategies, particularly rumination. Furthermore, longitudinal studies have been even less likely to assess the protective effects of more adaptive strategies. This study explored the longitudinal connections between depressivesymptoms and both potentially adaptive and maladaptive strategies. It found comparable prospective associations between adaptive and maladaptive strategies with symptoms. Additionally, stable moderate and high symptoms groups, as identified by latent class analysis, showed a greater tendency to use maladaptive strategies and a reduced tendency to use adaptive strategies. Thus, targeted interventions for youth with elevated depressive symptoms are needed to enhance their emotion regulation skill and to teach adaptive emotion regulation strategies, such as reappraisal, in order to broaden their strategies repertoire and reduce reliance on strategies like rumination, which tend to amplify emotional experiences. The results also underscore the importance of identifying specific factors that can alleviate the adverse effects of maladaptive emotion regulation strategies and facilitate the adoption of more effective strategies to mitigate the impact of everyday stressors in adolescents.